This project, now in the design phase, will assess whether or not it is possible to gather accurate information about an individual's past dietary intake. An answer to this question will permit researchers to evaluate retrospective data on the association between diet and cancer, and to plan further studies on this question. The study will make use of study subjects for whom dietary data was collected seven to fifteen years ago by means of seven-day diet records. These subjects will be interviewed, and their dietary intake during the past time period assessed by means of a food frequency questionnaire. Results will be evaluated with respect to the ability to classify individuals correctly, and the ability to determine the mean intake groups. In addition to improving the interpretation of observed diet-cancer associations, this technique, if validated, would permit the evaluation of an important potential confounder, past diet, in participants in chemoprevention trials.